MAJOR ARTICLE

Antibody Levels and Protection after Hepatitis B

Vaccine: Results of a 22-Year Follow-Up Study
and Response to a Booster Dose
Brian J. McMahon,1,2 Catherine M. Dentinger,2,a Dana Bruden,2 Carolyn Zanis,2 Helen Peters,2 Debbie Hurlburt,2
Lisa Bulkow,2 Anthony E. Fiore,3 Beth P. Bell,3 and Thomas W. Hennessy2
1
Liver Disease and Hepatitis Program, Alaska Native Tribal Health Consortium, and 2Arctic Investigations Program, Division of Emerging Infections
and Surveillance Services, National Center for Preparedness, Detection, and Control of Infectious Diseases, Centers for Disease Control
and Prevention (CDC), Anchorage, Alaska; 3Division of Viral Hepatitis, National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention,
CDC, Atlanta, Georgia

Background. The duration of protection in children and adults (including health care workers) resulting from
the hepatitis B vaccine primary series is unknown.
Methods. To determine the protection afforded by hepatitis B vaccine, Alaska Native persons who had received
plasma-derived hepatitis B vaccine when they were 16 months of age were tested for antibody to hepatitis B surface
antigen (anti-HBs) 22 years later. Those with levels !10 mIU/mL received 1 dose of recombinant hepatitis B
vaccine and were evaluated on the basis of anti-HBs measurements at 10–14 days, 30–60 days, and 1 year.
Results. Of 493 participants, 60% (298) had an anti-HBs level ⭓10 mIU/mL. A booster dose was administered
to 164 persons, and 77% responded with an anti-HBs level ⭓10 mIU/mL at 10–14 days, reaching 81% by 60
days. Response to a booster dose was positively correlated with younger age, peak anti-HBs response after primary
vaccination, and the presence of detectable anti-HBs before boosting. Considering persons with an anti-HBs level
⭓10 mIU/mL at 22 years and those who responded to the booster dose, protection was demonstrated in 87% of
the participants. No new acute or chronic hepatitis B virus infections were identified.
Conclusions. The protection afforded by primary immunization with plasma-derived hepatitis B vaccine during
childhood and adulthood lasts at least 22 years. Booster doses are not needed.

Hepatitis B vaccine (both plasma derived and recom-
binant) has been shown to be highly efficacious in pre-
venting infection with hepatitis B virus (HBV). Im-
munization with this vaccine starting at birth has
Received 15 April 2009; accepted 4 June 2009; electronically published 28
September 2009.
Potential conflicts of interest: none reported.
Presented in part: 43rd Annual Meeting of the Infectious Diseases Society of
America, San Francisco, 6–9 October 2005 (poster 1028).
Financial support: This study was funded in part through a cooperative
agreement (U50/CCU022279) between the Centers for Disease Control and
Prevention and the Alaska Native Tribal Health Consortium.
Disclaimer: The findings and conclusions in this report are those of the authors
and do not necessarily represent the views of Centers for Disease Control and
Prevention or the Alaska Native Tribal Health Consortium.
a
Present affiliation: Bureau of Communicable Diseases, New York City
Department of Health and Mental Hygiene, New York, New York.
Reprints or correspondence: Dr Brian J. McMahon, Liver Disease and Hepatitis
Program, Alaska Native Medical Center, 4315 Diplomacy Dr, Anchorage, AK 99508
(bdm9@cdc.gov).
The Journal of Infectious Diseases 2009; 200:1390–6
This article is in the public domain, and no copyright is claimed.
0022-1899/2009/20009-0006
DOI: 10.1086/606119
dramatically reduced the subsequent development of
chronic hepatitis B in young children from perinatal or
early childhood exposure to HBV [1, 2]. In US health
care workers, the incidence of acute HBV infection fell
substantially after Occupational Safety and Health Ad-
ministration requirements for vaccination were imple-
mented in the mid-1990s. However, the duration of
protection conferred by hepatitis B vaccination is in-
completely understood [3, 4]. Studies of infants im-
munized starting at birth suggest that some loss of im-
mune memory occurs with aging [5–7]. Less is known
about the persistence of antibody to hepatitis B surface
antigen (anti-HBs) and immune memory after vacci-
nation of older children and adults.
In 1981, immediately after licensure of hepatitis B
vaccine in the United States, we immunized a cohort
of 1578 Alaska Native adults and children 6 months or
older with 3 doses of plasma-derived hepatitis B vaccine
[8]. We performed yearly serologic testing for 11 years
and again at 15 years [4, 9, 10]. Fifteen years after their

1390 • JID 2009:200 (1 November) • McMahon et al

Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
 first dose of vaccine, 66% of participants had an anti-HBs level
⭓10 mIU/mL, the level thought to provide protection. No
participants had clinical signs or symptoms of acute symptom-
atic hepatitis, and none developed a chronic HBV infection.
Twenty-two years after the primary vaccination series, we
revisited a subset of this cohort to (1) determine the proportion
that had an anti-HBs level ⭓10 mIU/mL and (2) evaluate HBV
immune memory by administering a booster dose of hepatitis
B vaccine to those with an anti-HBs level !10 mIU/mL.
METHODS
Patients. Beginning in October 2003, we enrolled participants
from 7 of the 15 villages in the 1981 immunogenicity study
who had a documented response to a primary hepatitis B vac-
cination series. Persons from those 7 villages who had moved
to Anchorage or to the largest town in the region (Bethel,
Alaska) were also invited to participate. Persons in the 7 villages
did not differ from those in the other 8 villages in terms of
sex, age, initial response to the HBV vaccination series, or anti-
HBs level at the 15-year follow-up time point.
Persons who had received an additional dose of vaccine in
the intervening years were excluded. Persons in Anchorage were
recruited beginning in 2005.
Laboratory testing. Serologic specimens from participants
were tested for anti-HBs (ETI-AB-AUK PLUS and ABAU-STD-
SET; DiaSorin) at the Arctic Investigations Program in An-
chorage. The lower limit of detection of this assay was defined
as an anti-HBs value ⭓2 mIU/mL. Antibody to hepatitis B core
antigen (anti-HBc) was determined at the Alaska Native Med-
ical Center (Corzyme; Abbott Laboratories), and positive spec-
imens were further tested for hepatitis B surface antigen
(HBsAg) (Auszyme Monoclonal; Abbott Laboratories) and for
HBV DNA by nested polymerase chain reaction using primers
specific for the S gene, as described elsewhere [11]. An HBV
infection was defined as a positive anti-HBc or HBsAg result,
as described elsewhere [4, 9, 10].
Persons with an anti-HBs level ⭓10 mIU/mL were consid-
ered to be immune. Those with an anti-HBs level !10 mIU/
mL were offered an intramuscular booster dose of 10 mg of
hepatitis B vaccine (Recombivax HB; Merck). Anti-HBs con-
centration was determined at 10–14 days, 30–60 days, and 1
year after boosting. A positive booster dose response was de-
fined as an anti-HBs level rising to ⭓10 mIU/mL at either 10–
14 or 30–60 days. Vaccine recipients were given a 3-day diary
card to record adverse events.
The present study was approved by the Centers for Disease
Control and Prevention Institutional Review Board, the Alaska
Area Institutional Review Board, and 3 Alaska Native tribal
health boards: the Yukon-Kuskokwim Health Corporation, the
Alaska Native Tribal Health Consortium, and the Southcen-
Protection 22 Years after Hepatitis B Vaccination • JID 2009:200 (1 November) • 1391
Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
tral Foundation. All participants provided written informed
consent.
Hospital and village clinic records were reviewed for any
evidence of an illness in persons identified to be anti-HBc pos-
itive who were not already identified to be so in previous
surveys.
Statistical analysis. Qualitative anti-HBs levels are the pro-
portion of participants with ⭓10 mIU/mL and were compared
by the likelihood ratio x2 test for categorical covariates and by
the Cochran-Armitage test for trend for ordinal and continuous
covariates. Quantitative anti-HBs levels are presented as geo-
metric mean concentrations (GMCs). Anti-HBs levels are log
transformed; undetectable concentrations were assigned a value
of 1.0 mIU/mL before transformation. Anti-HBs levels (log
transformed) were compared between groups by the t test, the
paired t test, or analysis of variance, as appropriate. Exact con-
fidence intervals (CIs) for the incidence of new HBV infections
were calculated as derived by Garwood [12]. We compared the
time since loss of anti-HBs (time when the anti-HBs level de-
clined to !10 mIU/mL) between responders and nonresponders
to a booster dose of hepatitis B vaccine by the Wilcoxon rank
sum test. The estimated time of anti-HBs loss was calculated
as the chronological midpoint between the last anti-HBs level
⭓10 mIU/mL and the first anti-HBs level !10 mIU/mL.
We used logistic regression to test multivariate associations
with protective immunity at 22 years and with response to the
booster dose. The best-fit regression model was determined by
nonautomatic backward selection following the strategy of Hos-
mer and Lemeshow [13]. All 2-way interactions were evaluated
at P ! .05. For protective immunity at 22 years, we evaluated
age at primary vaccination series, sex, and anti-HBs level after
primary series. For response to a booster dose, we evaluated
age at booster dose, sex, anti-HBs level after primary series,
and preboost anti-HBs level. Age was treated as a categorical
variable in all statistical models. We used the Spearman rank
order statistic to assess colinearity among predictor variables.
Because the time since loss of anti-HBs was highly correlated
with anti-HBs level after the primary series and with preboost
anti-HBs level, it was not considered in the model. All statistical
analyses were conducted using SAS software (version 9.0; SAS),
and all P values are 2-sided. Statistical analysis was performed
by 2 of the authors (D.B. and L.B.).
RESULTS
Anti-HBs levels. Of 698 persons eligible, 493 (71%) partici-
pated (Figure 1). There were 455 participants residing in the
7 villages plus Bethel, and there were 38 in Anchorage. The
age, sex, and anti-HBs level after the primary vaccination series
were similar for participating persons (n p 493 ) and eligible
nonparticipating persons (n p 205) (data not shown).
The age, sex, and anti-HBs level on study enrollment are
 Figure 1. Participant flow chart in a 22-year follow-up study of 1578 persons receiving 3 doses of plasma-derived hepatitis B vaccine in Alaska in
1981. Anti-HBs, antibody to hepatitis B surface antigen; HBV, hepatitis B virus.

shown in Table 1. At the 22-year follow-up time point, the
GMC of anti-HBs among the 493 participants was 21.5 mIU/
mL; 298 (60%) had an anti-HBs level ⭓10 mIU/mL. Among
the 195 persons with an anti-HBs level !10 mIU, 138 (71%)
had detectable antibody (anti-HBs level between 2 and 9.9 mIU/
mL). In a multivariate logistic regression model, initial anti-
HBs level (P ! .001), age class (P ! .001), and sex (P p .04) were
associated with an anti-HBs level ⭓10 mIU/mL (Table 1).
Breakthrough HBV infections. Five participants (1%) had
a serum specimen positive for anti-HBc; none were positive
for HBsAg. All 5 were positive for anti-HBc on ⭓3 previous
consecutive evaluations; none were ever positive for HBV DNA.

1392 • JID 2009:200 (1 November) • McMahon et al

Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
 Table 1. Level of Antibody to Hepatitis B Surface Antigen (Anti-HBs) 22 Years after Hepatitis B Vaccination, by Sex, Age,
and Postvaccination Anti-HBs Level after the Primary Series—Alaska, 2003–2004

Participants with
Anti-HBs GMC, ⭓10 mIU/mL,
a
Characteristic No. (%) of cohort mIU/mL proportion (%) Univariate P Multivariate P
b
Sex .09 .04
Female 280 (57) 19.2 160/280 (57)
Male 213 (43) 25.0 138/213 (65)
Age at primary vaccination (age at .008b,c !.001
22-year follow-up)
0.5–4 (22–26) years 71 (14) 14.1 40/71 (56)
5–19 (27–41) years 268 (54) 29.1 179/268 (67)
20–49 (42–71) years 119 (24) 17.1 64/119 (54)
⭓50 (⭓72) years 35 (7) 11.1 15/35 (43)
d
Anti-HBs level after primary series !.001 !.001
10–99 mIU/mL 56 (11) 3.4 8/57 (14)
100–999 mIU/mL 168 (34) 11.1 64/168 (38)
1000–4999 mIU/mL 157 (32) 33.4 120/157 (76)
⭓5000 mIU/mL 111 (23) 239.9 106/111 (96)
All 493 (100) 21.5 298/493 (60)

NOTE. GMC, geometric mean concentration.

a
Logistic regression analysis.
b
Likelihood ratio x2 test.
c
For age 5–19 years versus all others, P p .002; for ⭓50 years of age versus all others, P p .03.
d
Cochran-Armitage test for trend.

Three persons previously positive for anti-HBc were no longer accounted for primarily by persons ⭓60 years of age, because
positive at this follow-up time point. The cumulative incidence this age group had a lower percentage of participants with an
of HBV infections over the course of 22 years of follow-up was anti-HBs level ⭓10 mIU/mL (P p .004) and a significantly
0.74 (95% CI, 0.31–1.45) cases per 1000 persons per year. lower anti-HBs GMC at 10–14 days (P ! .001) and at 30–60
Response to a booster dose of hepatitis B vaccine. Of 195 days (P p .006) compared with all other groups combined (Ta-
eligible persons with a baseline anti-HBs level !10 mIU/mL, ble 3). The response to the booster dose at all time points was
165 (85%) received a booster dose of hepatitis B vaccine (Fig- positively correlated with the participants’ response to the pri-
ure 1). The median age of the 164 persons who provided a se- mary vaccination series (Table 3).
rum specimen after the booster dose was 37 years (range, 22– Those participants who had a preboost anti-HBs level 2–9.9
83 years); 99 (60%) were female. mIU/mL had a significantly higher probability of achieving a
Responses to the booster dose are shown in Table 2. Of the booster response than did those with no measurable anti-HBs
155 tested at 10–14 days and 30–60 days, 120 (77%) responded
with an anti-HBs level ⭓10 mIU/mL. If 13 persons who did Table 2. Level of Antibody to Hepatitis B Surface Antigen (Anti-
not respond or were not tested at 10–14 days but who re- HBs) after a Booster Dose of Hepatitis B Vaccine among Partic-
sponded at 30–60 days are included, then 133 (81%) of 164 ipants with an Initial Anti-HBs Level !10 mIU/mL—Alaska, 2003–
2004
had a protective anti-HBs level measured 10–60 days after the
booster dose. The GMC for the group measured 30–60 days Time after booster dose
after the booster dose was significantly lower than that mea-
10–14 days 30–60 days 1 year
sured 6 months after the primary vaccination series adminis- Characteristic (n p 155) (n p 163) (n p 155)
tered in 1981 (60.6 vs 280.9 mIU/mL; P ! .001 ) [8]. Of the 155
Anti-HBs GMC, mIU/mL 91.1 60.6 8.0
participants tested 1 year after the booster dose, 63 (41%) had Anti-HBs category
an anti-HBs level ⭓10 mIU/mL; the GMC had fallen to 8.0 !10 mIU/mL 35 (23) 38 (23) 92 (59)
mIU/mL. 10–49 mIU/mL 21 (14) 36 (22) 50 (32)
We evaluated results from the 145 persons who had serum 50–499 mIU/mL 44 (28) 57 (35) 8 (5)
drawn at all 3 follow-up time points (Table 3). At 10–14 days, ⭓500 mIU/mL 55 (35) 32 (20) 5 (3)
there was a difference in the proportion with an anti-HBs level NOTE. Data are no. (%) of participants, unless otherwise indicated. GMC,
⭓10 mIU/mL by age (P p .01) (Table 3). This difference was geometric mean concentration.

Protection 22 Years after Hepatitis B Vaccination • JID 2009:200 (1 November) • 1393

Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
 Table 3. Level of Antibody to Hepatitis B Surface Antigen (Anti-HBs) after a Booster Dose of Hepatitis B Vaccine in 145 Persons
with an Anti-HBs Level !10 mIU/mL 22 Years after a Primary Hepatitis B Vaccination Series Who Had Anti-HBs Levels Measured at
All 3 Follow-up Time Points—Alaska, 2003–2004

Time since booster dose

10–14 days 30–60 days (visit 2) 1 year

GMC, % with ⭓10 GMC, % with ⭓10 GMC, % with ⭓10

a a a
Characteristic mIU/mL mIU/mL P mIU/mL mIU/mL P mIU/mL mIU/mL P
Sex .07 .45 .63
Female (n p 90) 127.3 82 78.5 80 8.6 42
Male (n p 55) 51.9 69 51.9 75 7.6 38
Age at booster vaccination .01 .52 .23
!30 years (n p 35) 108.4 74 72.9 80 7.7 34
30–39 years (n p 46) 207.4 83 125.1 80 10.1 50
40–59 years (n p 33) 100.4 91 64.6 82 9.7 45
⭓60 years (n p 31) 19.4 58 24.8 68 5.4 29
Anti-HBs level measured after primary series !.001 !.001 !.001
10–99 mIU/mL (n p 39) 12.3 54 14.9 62 3.7 13
100–499 mIU/mL (n p 49) 56.0 76 42.1 69 9.0 43
500–999 mIU/mL (n p 23) 168.1 91 122.6 96 9.0 48
⭓1000 mIU/mL (n p 34) 1174.8 97 482.6 97 17.1 65
Preboost anti-HBs level !.001 .002 !.001
!2 mIU/mL (n p 42) 21.1 55 22.1 60 3.7 14
2–4 mIU/mL (n p 49) 76.4 82 61.0 80 10.8 49
5–9 mIU/mL (n p 54) 329.1 91 171.8 91 11.8 55
Time since loss of anti-HBs level ⭓10 mIU/mL !.001 !.001 !.001
!7 years (n p 35) 624.7 100 275.5 97 14.4 57
7–11 years (n p 34) 142.8 85 92.8 88 10.6 47
12–16 years (n p 40) 106.2 80 81.3 85 9.2 45
⭓17 years (n p 44) 7.6 44 9.9 42 3.3 14
All (n p 145) 90.6 77 66.8 78 8.2 41

NOTE. GMC, geometric mean concentration.

a
P values are for comparison of the percentages of participants with an anti-HBs level ⭓10 mIU/mL.

level (P ! .001 at all 3 follow-up time points). Persons with a
longer time since the loss of a protective anti-HBs level were
less likely to respond to a booster dose than were those with
a more recent loss (Table 3). The proportion of participants
responding to the booster dose (as evidenced by an anti-HBs
level ⭓10 mIU/mL) did not differ by village and ranged from
76% (13/17) to 86% (19/22). At the time of the original vaccine
project in 1981, the prevalence of persons positive for HBsAg
in these villages ranged from 1.2% to 8.6%; at the time of the
22-year follow-up, the prevalence had fallen to a range of
0.3%–5.3%.
In a multivariate model, we found that the response to a
booster dose was significantly associated with anti-HBs level
after the primary vaccination series (P ! .001 ) and age at the
time of the booster dose (P p .02). The predicted probabilities
of responding to a booster dose 22 years after the primary
series are displayed in Figure 2. There was no difference in the
proportion who responded between those !30 years of age and
those 30–39 years of age; therefore, we combined these 2 age
classes. Similarly, there was no difference between participants
with an anti-HBs level of 500–999 and ⭓1000 mIU/mL after
the primary series, so we also combined these 2 groups. Par-
ticipants aged 40–59 years and those who had an anti-HBs level
⭓500 mIU/mL after the primary series were most likely to
respond to the booster dose (Figure 2). No serious adverse
events to the booster dose were reported.
Of the 493 study participants, 431 (87% [95% CI, 84.5%–
90.3%]) demonstrated long-term protection from hepatitis B
vaccination, as defined by an anti-HBs level ⭓10 mIU/mL at
the screening visit or by a response to a booster dose of ⭓10
mIU/mL (Figure 1). Excluding persons who had a level !10
mIU/mL at the initial screening visit but who did not receive
a booster dose (n p 31), a total of 93% (95% CI, 91.0%–95.6%)
of the cohort was protected. By applying the results of the
multivariate logistic regression model to these 31 persons strat-
ified by age and primary response level, we estimate that 25
(80%) would have responded to the booster dose if they had
participated. Combining the observed response in 431 of 462
participants with the predicted response in 25 of 31 partici-
pants, we estimate that 92.5% of the cohort was protected.

1394 • JID 2009:200 (1 November) • McMahon et al

Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018

Figure 2. Predicted percentage of persons responding to a hepatitis
B vaccine booster dose (antibody to hepatitis B surface antigen [anti-
HBs] level 110 mIU/mL), by age and anti-HBs level after the primary
vaccination series—Alaska, 2003–2004. Error bars indicate 95% confi-
dence intervals from logistic regression.
DISCUSSION
We report here findings from the largest and longest follow-
up study to date of children and adults who responded to
primary vaccination with plasma-derived hepatitis B vaccine.
After 22 years, an estimated 92.5% of primary responders had
evidence of continued protection (anti-HBs level ⭓10 mIU/
mL or humoral immune memory defined by response to a
booster dose). Furthermore, no participants became chronically
infected, had laboratory evidence of acute hepatitis B, or had
detectable HBV DNA.
When this cohort was recruited in 1981, HBV infection was
hyperendemic among Alaska Native people in this region, with
a prevalence of HBsAg of 8.2% (range, 0–27.5% per village)
[8]. Since then, the introduction of hepatitis B vaccination as
a routine childhood immunization coupled with a large-scale
catch-up program involving all age groups has virtually elim-
inated new infections in Alaska. The incidence of acute symp-
tomatic hepatitis B has fallen from 1200 cases per 100,000
population in 1981 to 0 cases per 100,000 population since
1994 (B.J.M., unpublished data). In addition, 198% of chron-
ically infected persons from this area have seroconverted from
hepatitis B e antigen to antibody to hepatitis B e antigen and
therefore are much less infective [14]. During the first 10 years
of follow-up, we found that only 8% of participants had evi-
dence of an increase in anti-HBs level without concomitant
anti-HBc during follow-up, suggesting that natural exposure to
HBV was uncommon [15]. Thus, we believe our findings are
likely applicable to populations with either high or low endem-
icity of HBV infection.
A larger proportion of participants who received the primary
vaccination series at 5–19 years of age had an anti-HBs level
⭓10 mIU/mL 22 years later, compared with other age groups.
Protection 22 Years after Hepatitis B Vaccination • JID 2009:200 (1 November) • 1395
Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
Participants who were vaccinated during older childhood or
adolescence are similar in age to children who were first vac-
cinated during catch-up hepatitis B vaccination programs in
the United States and elsewhere [2]. Our study suggests that
children vaccinated through these catch-up programs are likely
to be protected from HBV infection for decades.
Among participants with an anti-HBs level !10 mIU/mL, we
found that the probability of boosting was positively associated
with the presence of detectable anti-HBs (2–9.9 mIU/mL) and
was negatively associated with a documented anti-HBs level
!10 mIU/mL for 17 years. This suggests that the ability to
mount an anamnestic response to a booster dose of hepatitis
B vaccine may wane in those who do not have detectable anti-
HBs for prolonged periods of time. No difference was found
in the proportion of participants who had a booster response
by village, despite a wide variation in the prevalence of persons
positive for HBsAg.
While the protective level of anti-HBs after primary vacci-
nation needed to prevent HBV infection has been shown to be
10 mIU/mL in controlled clinical trials [16, 17], the level of
antibody necessary to provide long-term protect against HBV
infection is unknown. It is likely that humoral antibody con-
centration might not be the best indicator of long-term pro-
tection, because persons might be protected against chronic
infection by other immune mechanisms, such as cellular im-
munity. Because HBV infection has such a long incubation
period (average, 60–90 days), there might be time for residual
cellular immunity to prevent clinical illness and chronic infec-
tion. The appearance of anti-HBc in a small minority in this
cohort might indicate such breakthrough infections in persons
who retain sufficient cellular immunity to prevent symptomatic
or chronic infection. Assays to measure correlates of protection
many years after vaccination are needed.
Previous studies of anti-HBs persistence after hepatitis B
vaccination of children and adults from regions of both high
and low endemicity have shown excellent protection 4–10 years
after immunization, similar to the results through the first 10
years of follow-up of this Alaska cohort [4, 18–21]. The few
published studies following subjects after initial hepatitis B vac-
cination beyond 10 years did not evaluate response to a booster
dose but found persistence of anti-HBs up to 18 years [6, 22,
23], similar to our study.
Follow-up studies of children vaccinated against hepatitis B
beginning at birth have demonstrated the presence of protective
anti-HBs levels at 14 years of age in !10% of those who received
a 3-dose schedule and in only 37% of those who received a 4-
dose schedule. This is a more rapid decline in anti-HBs level
compared with that observed in studies of persons vaccinated
as older children or adults [3, 5–7, 24]. In addition, 29%–50%
of adolescents immunized at birth did not demonstrate an
 anamnestic response to a booster dose [24, 25]. A difference
in the duration of immunity among those vaccinated beginning
at birth compared with older children and adults could be a
consequence of a less-mature immune system at the time of
primary immunization. Further evaluation of long-term pro-
tection among cohorts vaccinated beginning at birth should be
a high priority [2].
A potential limitation of the generalizability of our findings
deserves mention. Participants in our study received the pri-
mary hepatitis B vaccination series with the plasma-derived
vaccine, which is no longer used in the United States. Antibody
response and vaccine effectiveness in this and other populations
that received the plasma-derived vaccine are similar to those
demonstrated in populations vaccinated with recombinant vac-
cines beginning in the late 1980s [2]. For these reasons, we
believe our findings are applicable to persons who received
either vaccine type. However, the validity of extending our
findings to persons immunized with recombinant hepatitis B
vaccines needs to be confirmed by long-term studies similar to
ours.
In conclusion, we believe the results of our study are directly
relevant to health care workers, including those vaccinated with
plasma-derived vaccine in the early 1980s, and to children and
adolescents who were vaccinated during catch-up hepatitis B
vaccination programs in the 1990s [2]. In light of the strong
evidence we present here, hepatitis B vaccine booster doses are
not currently indicated. Our unique longitudinal data are cru-
cial for the ongoing evaluation of public health vaccination
strategies and programs. For these reasons, we plan additional
follow-up of this cohort 30 years after the primary vaccination
series.
Acknowledgments
We thank the community health practitioners and aides from the villages
involved in the study for assisting us with the fieldwork, the Yukon Kus-
kokwim Health Corporation for its support, and Stephanie Bialek, James
Gove, Peggy McMahon, Priti Patel, and Colin Shepard, who helped with
the fieldwork.
References
1. Mast EE, Alter MJ, Margolis HS. Strategies to prevent and control
hepatitis B and C virus infections: a global perspective. Vaccine 1999;17:
1730–3.
2. Mast EE, Margolis HS, Fiore AE, et al. A comprehensive immunization
strategy to eliminate transmission of hepatitis B virus infection in the
United States: recommendations of the Advisory Committee on Im-
munization Practices (ACIP) part 1: immunization of infants, children,
and adolescents. MMWR Recomm Rep 2005; 54:1–31.
3. Dentinger CM, McMahon BJ, Butler JC, et al. Persistence of antibody
to hepatitis B and protection from disease among Alaska natives im-
munized at birth. Pediatr Infect Dis J 2005; 24:786–92.
4. McMahon BJ, Bruden DL, Petersen KM, et al. Antibody levels and
1396 • JID 2009:200 (1 November) • McMahon et al
Downloaded from https://academic.oup.com/jid/article-abstract/200/9/1390/851238
by guest
on 04 March 2018
protection after hepatitis B vaccination: results of a 15-year follow-up.
Ann Intern Med 2005; 142:333–41.
5. Lu CY, Chiang BL, Chi WK, et al. Waning immunity to plasma-derived
hepatitis B vaccine and the need for boosters 15 years after neonatal
vaccination. Hepatology 2004; 40:1415–20.
6. van der Sande MAB, Waight P, Mendy M, et al. Long-term protection
against carriage of hepatitis B virus after infant vaccination. J Infect
Dis 2006; 193:1528–35.
7. Petersen KM, Bulkow LR, McMahon BJ, et al. Duration of hepatitis
B immunity in low risk children receiving hepatitis B vaccinations from
birth. Pediatr Infect Dis J 2004; 23:650–5.
8. Heyward WL, Bender TR, McMahon BJ, et al. The control of hepatitis-
B-infection with vaccine in Yupik Eskimos—demonstration of safety,
immunogenicity, and efficacy under field conditions. Am J Epidemiol
1985; 121:914–23.
9. Wainwright RB, Bulkow LR, Parkinson AJ, Zanis C, McMahon BJ.
Protection provided by hepatitis B vaccine in a Yupik Eskimo popu-
lation—results of a 10-year study. J Infect Dis 1997; 175:674–7.
10. Wainwright RB, McMahon BJ, Bulkow LR, et al. Duration of im-
munogenicity an efficacy of hepatitis-B vaccine in a Yupik Eskimo
population. JAMA 1989; 261:2362–6.
11. Livingston SE, Simonetti JP, McMahon BJ, et al. Hepatitis B virus
genotypes in Alaska native people with hepatocellular carcinoma: pre-
ponderance of genotype F. J Infect Dis 2007; 195:5–11.
12. Garwood F. Fiducial limits for the poisson distribution. Biometrika
1936; 46:441–53.
13. Hosmer DW, Lemeshow S. Applied logistic regression. 2nd ed. New
York: John Wiley & Sons, 2000.
14. Livingston SE, Simonetti JP, Bulkow LR, et al. Clearance of hepatitis
B e antigen in patients with chronic hepatitis B and genotypes A, B,
C, D, and F. Gastroenterology 2007; 133:1452–7.
15. Bulkow LR, Wainwright RB, McMahon BJ, Parkinson AJ. Increases in
levels of antibody to hepatitis B surface antigen in an immunized
population. Clin Infect Dis 1998; 26:933–7.
16. Szmuness W, Stevens CE, Harley EJ, et al. Hepatitis-B vaccine—dem-
onstration of efficacy in a controlled clinical-trial in a high-risk pop-
ulation in the United-States. N Engl J Med 1980; 303:833–41.
17. Francis DP, Hadler SC, Thompson SE, et al. The prevention of hep-
atitis-B with vaccine: report of the Centers for Disease Control multi-
center efficacy trial among homosexual men. Ann Intern Med 1982;
97:362–6.
18. Barash C, Conn MI, DiMarino AJ, Marzano J, Allen ML. Serologic
hepatitis B immunity in vaccinated health care workers. Arch Intern
Med 1999; 159:1481–3.
19. Milne A, Hopkirk N, Moyes CD. Hepatitis-B vaccination in children—
persistence of immunity at 9 years. J Med Virol 1994; 44:113–4.
20. Pasko MT, Beam TR. Persistence of anti-HBs among health-care per-
sonnel immunized with hepatitis-B vaccine. Am J Public Health 1990;
80:590–3.
21. Tabor E, Cairns J, Gerety RJ, Bayley AC. Nine-year follow-up study
of a plasma-derived hepatitis B vaccine in a rural African setting. J
Med Virol 1993; 40:204–9.
22. Williams JL, Christensen CJ, McMahon BJ, et al. Evaluation of the
response to a booster dose of hepatitis B vaccine in previously im-
munized healthcare workers. Vaccine 2001; 19:4081–5.
23. Yuen MF, Lim WL, Chan AO, Wong DK, Sum SS, Lai CL. 18-year
follow-up study of a prospective randomized trial of hepatitis B vac-
cinations without booster doses in children. Clin Gastroenterol Hepa-
tol 2004; 2:941–5.
24. Lu CY, Ni YH, Chiang BL, et al. Humoral and cellular immune re-
sponses to a hepatitis B vaccine booster 15–18 years after neonatal
immunization. J Infect Dis 2008; 197:1419–26.
25. Samandari T, Fiore AE, Negus S, et al. Differences in response to
hepatitis B vaccine booster dose among Alaskan children and adoles-
cents vaccinated during infancy. Pediatrics 2007; 120:e373–81.